Field of the Invention
The present invention is related to a data transmitting method, and more particularly to a method that employs a switch to transmit data between a host and a storage without data passing through a mother board of the switch such that data can be transmitted much more effectively than before.
Description of the Prior Art
Since Peripheral Component Interconnect Express (PCIe) is one of the commonly used PC system bus protocols, it can be applied to PC interconnections or to other external IC chips, for example, a graphic card with graphic processing units. Alternatively, PCIe protocols can also be utilized for connecting a PC to its I/O devices such that data, for example, multimedia info, can be transmitted there-in-between. In general, a data transmitting speed of PCIe protocol is about 500 MBps to 8 GBps depending on different types of bus widths or specifications. Recently, for speeding up the data transmitting efficiency and achieving full-duplex transmission, a brand new protocol known as Thunderbolt (TB) has been provided.
Specifically, TB is a novel protocol used for full-duplex transmission in high speed I/O technology. Its efficiency is beyond that of all I/O technologies nowadays and can increase the bi-directional transmitting speed by more than 10 GBps. Moreover, TB protocol merely utilized a single wire to transmitting both data for PCIe and Display Port at the same time, which in turns provides much more flexibility and simplicity in practical than ever. However, when TB protocols are applied nowadays, problems still arise. For example, when there are a plurality of storages which are commonly connected to a plurality of hosts, then the command sent from the host must be transformed through a certain process to become a certain format such that the command can be transmitted from the host to the storage. Under such circumstance, since the command sent from the host must be transformed through a certain process to become a certain format, only the storage which can read the certain format can be adopted in the structure. Therefore, when applying this method in practical, the types of storages which can be used are seriously limited, and lacks of universality too.
Another prior method was purposed, in which the command sent from the host is transmitted to an intermediate mother board first, and the data to be sent is accompanied by the command to be transmitted to the intermediate mother board, too. Later, the command is transformed into a universal command and is transmitted to the storage. Meanwhile, the data to be sent is also transmitted to the storage for the storage to access. The benefits of applying this method are that, the command eventually got to the storage is the universal format that most of the storages can access. Therefore, most of the storages nowadays which can read the universal format can be adopted in the system structure, solving the above mentioned problems occurred in the previous method. Nevertheless, what causes the more severe problems are when using this method, both the command and the data to be sent are transmitted through 1. the host to the intermediate mother board; and 2. the intermediate mother board to the storage. As a result, a much more high-level mother board is a must in the structure in order to handle such a great amount of computation. In addition, the space for the memory inside the mother board (i.e. DRAM) has to accordingly increase as well.
Moreover, if the numbers of hosts and storages connected thereto are increased, then the mother board and its memory space inside have to be much more advanced and sophisticated too. In other words, the cost consumption is a terrible issue to overcome.
On account of all, it should be obvious that there is indeed an urgent need for the professionals in the field for a new data transmitting method to be developed that can actively and aggressively induce the data to be transmitted between at least one host and storage so as to solve the above-mentioned problems occurring in the prior design.